Research On Health State Prediction Method Of Aero-engine Gas Path System Based On Multi-features

Aero-engine is the power system of aircraft,and its health states are key to the safe and reliable operation of aircraft.Among the auxiliary systems of aero-engines,the system with the highest failure rate is the gas path system,and the study of the health state prediction for the gas path system is the key link of aero-engine health management.Due to the complexity of the gas path system,it is difficult to establish a dynamic and comprehensive health state prediction model.Therefore,it is necessary to use various monitoring data,expert knowledge and other information of the aero-engine gas path system to reflect the health states for the system.Under the framework of belief rule base(BRB)theory,this paper uses multiple performance parameters of the gas path system as input features,integrates expert knowledge,and studies the health state prediction method for aero-engine gas path system based on multi-feature,which comprehensively reflects the health states of the aero-engine gas path system.First,the working mechanism and failure mechanism of the aero-engine gas path system are analyzed and the health state feature quantity are determined.A health state prediction model for aero-engine gas path system based on multi-feature belief rule base(MBRB)is established.Compared with the traditional prediction of the a single or two performance parameters of gas path systems,the proposed method contains more abundant health state information.At the same time,the projection covariance matrix adaptation evolutionary strategies(P-CMA-ES)be used to updates the model parameters,which overcome the subjectivity of expert knowledge in the initial belief rule base.Secondly,considering the harshness of the operating environment of the aero-engine gas path system,the monitoring data has the problems of environmental interference and sensor performance degradation.A health state prediction for the aero-engine gas path system based on multi-feature belief rule base with considering monitoring error(MBRB-?)is established.Thirdly,considering the huge model structure will reduce the real-time performance of the model in engineering practice,and the single model structure will reduce the prediction effect of the model.The health states of the gas path system cannot be fully reflected.A health state prediction model for aero-engine gas path system based on parallel-serial belief rule base is established,which solves the problems of high complexity and long running time of a single belief rule base model.And fuzzy c-means is used to reduce the uncertainty of expert knowledge.Finally,the validity of the health state prediction model for the aero-engine gas path system is proved by using the monitoring data of a certain type of aero-engine.Furthermore,based on the health state prediction model proposed in this paper,a visual health state prediction software of aero-engine gas path system is developed,which verifies the effectiveness and accuracy of the method proposed in this paper.